<\/p>\n
R’de kar\u015f\u0131la\u015fabilece\u011finiz bir hata:<\/span><\/p>\n Bu hata genellikle bir regresyon modelinde \u00e7oklu do\u011frusall\u0131k<\/a> mevcut oldu\u011funda ortaya \u00e7\u0131kar. Yani, modeldeki iki veya daha fazla yorday\u0131c\u0131 de\u011fi\u015fken y\u00fcksek oranda (veya m\u00fckemmel \u015fekilde) ili\u015fkilidir.<\/span><\/p>\n Bu oldu\u011funda, bir de\u011fi\u015fkenin ba\u015fka bir de\u011fi\u015fkenin “takma ad\u0131” oldu\u011funu s\u00f6yleriz, bu da regresyon modelinin yerle\u015ftirilmesinde sorunlara neden olur.<\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnekte bu hatan\u0131n pratikte nas\u0131l d\u00fczeltilece\u011fi g\u00f6sterilmektedir.<\/span><\/p>\n R’de a\u015fa\u011f\u0131daki regresyon modelini<\/a> uygulad\u0131\u011f\u0131m\u0131z\u0131 varsayal\u0131m:<\/span><\/p>\n \u00c7oklu ba\u011flant\u0131n\u0131n bir sorun olup olmad\u0131\u011f\u0131n\u0131 belirlemek amac\u0131yla modeldeki her bir tahmin de\u011fi\u015fkeninin VIF de\u011ferlerini hesaplamak i\u00e7in araba<\/strong> paketinin vive()<\/strong> fonksiyonunu kullanabiliriz:<\/span><\/p>\n ” Modelde takma katsay\u0131lar var” hatas\u0131 al\u0131yoruz.<\/strong> \u201c<\/span><\/p>\n Bu bize modeldeki iki veya daha fazla yorday\u0131c\u0131 de\u011fi\u015fkenin m\u00fckemmel bir \u015fekilde ili\u015fkili oldu\u011funu s\u00f6yler.<\/span><\/p>\n Hangi yorday\u0131c\u0131 de\u011fi\u015fkenlerin m\u00fckemmel \u015fekilde korelasyona sahip oldu\u011funu belirlemek amac\u0131yla de\u011fi\u015fkenler i\u00e7in birkorelasyon matrisi<\/a> olu\u015fturmak amac\u0131yla cor()<\/strong> fonksiyonunu kullanabiliriz:<\/span><\/p>\n x2<\/strong> ve x3<\/strong> de\u011fi\u015fkenlerinin korelasyon katsay\u0131s\u0131n\u0131n<\/a> 1 oldu\u011funu g\u00f6rebiliriz. Bu bize bu iki de\u011fi\u015fkenin m\u00fckemmel korelasyona sahip olduklar\u0131 i\u00e7in hataya neden oldu\u011funu s\u00f6yler.<\/span><\/p>\n Bu hatay\u0131 d\u00fczeltmek i\u00e7in regresyon modelini yeniden ayarlaman\u0131z ve bu iki de\u011fi\u015fkenden birini d\u0131\u015far\u0131da b\u0131rakman\u0131z yeterlidir.<\/span><\/p>\n Her ikisi de regresyon modelinde tamamen ayn\u0131 bilgiyi sa\u011flad\u0131\u011f\u0131 i\u00e7in hangi de\u011fi\u015fkeni atlad\u0131\u011f\u0131m\u0131z \u00f6nemli de\u011fil.<\/span><\/p>\n Basit olmas\u0131 a\u00e7\u0131s\u0131ndan x3’\u00fc<\/strong> \u00e7\u0131karal\u0131m ve regresyon modelini tekrar yerle\u015ftirelim:<\/span><\/p>\n Modelin VIF de\u011ferlerini hesaplarken bu sefer herhangi bir hata almad\u0131\u011f\u0131m\u0131z\u0131 unutmay\u0131n \u00e7\u00fcnk\u00fc \u00e7oklu ba\u011flant\u0131 art\u0131k sorun de\u011fil.<\/span><\/p>\n \u0130lgili:<\/strong> R’de VIF De\u011ferleri Nas\u0131l Hesaplan\u0131r ve Yorumlan\u0131r?<\/a><\/span><\/p>\n A\u015fa\u011f\u0131daki e\u011fitimlerde R’deki di\u011fer yayg\u0131n hatalar\u0131n nas\u0131l d\u00fczeltilece\u011fi a\u00e7\u0131klanmaktad\u0131r:<\/span><\/p>\n R’de nas\u0131l onar\u0131l\u0131r: de\u011fi\u015ftirmenin uzunlu\u011fu s\u0131f\u0131rd\u0131r<\/a> R’de kar\u015f\u0131la\u015fabilece\u011finiz bir hata: Error in vive.default(model): there are aliased coefficients in the model Bu hata genellikle bir regresyon modelinde \u00e7oklu do\u011frusall\u0131k mevcut oldu\u011funda ortaya \u00e7\u0131kar. Yani, modeldeki iki veya daha fazla yorday\u0131c\u0131 de\u011fi\u015fken y\u00fcksek oranda (veya m\u00fckemmel \u015fekilde) ili\u015fkilidir. Bu oldu\u011funda, bir de\u011fi\u015fkenin ba\u015fka bir de\u011fi\u015fkenin “takma ad\u0131” oldu\u011funu s\u00f6yleriz, bu da regresyon modelinin […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[11],"tags":[],"class_list":["post-2330","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n Error in vive.default(model): there are aliased coefficients in the model\n<\/strong><\/pre>\n
Hata nas\u0131l yeniden olu\u015fturulur?<\/strong><\/span><\/h3>\n
#make this example reproducible\n<\/span>set. seeds<\/span> (0)\n\n#define data\n<\/span>x1 <- rnorm(100)\nx2 <- rnorm(100)\nx3 <- x2*3\ny <- rnorm(100)\n\n#fit regression model\n<\/span>model <- lm(y~x1+x2+x3)\n<\/strong><\/span><\/pre>\n library<\/span> (car)\n<\/span><\/span>\n#calculate VIF values for predictor variables<\/span>\nlively(model)\n\n<\/span>Error in vive.default(model): there are aliased coefficients in the model\n<\/strong><\/span><\/pre>\n Hata nas\u0131l d\u00fczeltilir?<\/strong><\/span><\/h3>\n
#place variables in data frame\n<\/span>df <- data. frame<\/span> (x1, x2, x3, y)\n\n#create correlation matrix for data frame\n<\/span>cor(df)\n\n x1 x2 x3 y\nx1 1.00000000 0.126886263 0.126886263 0.065047543\nx2 0.12688626 1.000000000 1.000000000 -0.009107573\nx3 0.12688626 1.000000000 1.000000000 -0.009107573\ny 0.06504754 -0.009107573 -0.009107573 1.000000000\n<\/span><\/span><\/strong><\/pre>\n library<\/span> (car)\n<\/span>\n#make this example reproducible\nset. seeds<\/span> (0)\n\n#define data\n<\/span>x1 <- rnorm(100)\nx2 <- rnorm(100)\nx3 <- x2*3\ny <- rnorm(100)\n\n#fit regression model\n<\/span>model <- lm(y~x1+x2)\n\n#calculate VIF values for predictor variables in model<\/span>\nlively(model)\n\n x1 x2 \n1.016364 1.016364 \n<\/span><\/span><\/strong><\/pre>\n Ek kaynaklar<\/strong><\/span><\/h3>\n
R’de nas\u0131l d\u00fczeltilir: ba\u011f\u0131ms\u0131z de\u011fi\u015fkenler farkl\u0131 say\u0131da sat\u0131r i\u00e7erir<\/a>
R’de nas\u0131l d\u00fczeltilir: arg\u00fcman ne say\u0131sal ne de mant\u0131ksal: return na<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"